[Prism54-devel] Load firmware @ probe

[Aurelien]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,
I see several drawbacks about loading the firmware at probe time :
~  - security : while not configured, your device will be working and
trying to assoc or doing other things.
~  - energy loss.
~  - it works just fine currently and this single line might have big
consequences.
~  - hvr designed all the booting process and as far as i remember, he
had several good objections against this behavior.

The only advantage I can see is knowing the MAC address of the card
earlier, which I do not think is really significant.

	Aurelien

Luis R. Rodriguez wrote:
| OK, I'm stubborn, what can I say. I'd personally would prefer the driver
| load the firmware @ probe time. Attached patch does this. This also
| allows for the re-loading of the firmware @ open time. We previously
| hadn't done this because -- what if the driver was built-in. But you
| know what -- if you have the driver built-in right now you're still
| screwed if you don't have the firmware available @ bootup.
|
| Before I make this change I'd just like to ask *again* for comments in
| favor/against enabling this behaviour.
|
| 	Luis
|
|
|
| ------------------------------------------------------------------------
|
| diff -Naur prism54-cvs-latest/ksrc/islpci_hotplug.c
prism54-cvs-latest-mcgrof/ksrc/islpci_hotplug.c
| --- prism54-cvs-latest/ksrc/islpci_hotplug.c	2004-03-20
11:58:36.000000000 -0500
| +++ prism54-cvs-latest-mcgrof/ksrc/islpci_hotplug.c	2004-04-14
19:01:52.000000000 -0400
| @@ -299,6 +299,7 @@
|
|  	/* Pretty card model discovery output */
|  	prism54_get_card_model(ndev);
| +	islpci_reset(priv, 1);
|
|  	return 0;
|

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFAgP4tQ9u0bSbsoIsRAqhZAJwOmz2ht7xrPZd1gSF7GMajcOt5XACgnshn
PjIwm3we0SU8ZD1QJtOoFnw=
=Swya
-----END PGP SIGNATURE-----